Transition to Plants of the World Online as a taxonomic reference

iNaturalist used to use several regional floras as vascular plant taxonomic references with The Plant List as the tie-breaker. As discussed here, we are now using Plants of the World Online (POWO) as the primary vascular plant taxonomic reference with two exceptions:
1) POWO isn't yet complete for genera beginning with the letter 'P' (e.g. Phacelia) or Ferns (Class Polypodiopsida).
2) We maintain a list of potential deviations from POWO that we will discuss on a case by case basis. Those we determine to be 'Accepted Deviations' take precedence over POWO.

If you'd like to propose a deviation from POWO, flag the taxon (if its not already flagged) kick off a discussion for why you think the name is valid despite not being in POWO and also add the name to the list of potential deviations.

Here's a decision tree for how to go about helping resolve vascular plant species in iNaturalist that are not in POWO:

Names added to the potential deviation list will be discussed (in their corresponding flag) and potentially discussed with Kew Botanical Garden staff (which maintains POWO). From there we will decide whether to go with POWO or maintain them as 'Accepted Deviations'. As POWO updates, any corresponding rows should be removed from the potential deviation spreadsheet.

For example, Lithospermum parviflorum is an active species in iNaturalist not valid in POWO. It was added to the potential deviation list while @bouteloua led a discussion that seemed to indicate that it should be valid in POWO. This recommendation was brought to the attention of Kew and they agree it should be a valid species. We responded by marking it as an 'Accepted Deviation'. If Kew updates POWO by adding Lithospermum parviflorum, it should be removed from the list (as it will be no longer needed).

Posted by loarie loarie, August 20, 2018 11:34 PM

Comments

Thumb

Exciting. :)

Was there any comment on whether clearer information on taxon splits would be incorporated into POWO?

Posted by bouteloua 7 months ago (Flag)
Thumb

Can we talk about how POWO frequently fails to list nominate vars/subsps in any given parent taxon's "Accepted Infraspecifics" section?

For example, Cirsium hydrophilum, which is (in reality) parent to both a nominate variety and also to C. h. var. vaseyi: http://plantsoftheworldonline.org/taxon/urn:lsid:ipni.org:names:59046-2

Only the latter variety is listed, even as the former is federally endangered! I immediately was able to find example after example of these omissions (Lasthenia californica ssp. californica was the lone counterexample I found in a few minutes of searching).

If we intend to mark deviations, these are going to surface over and over again.

Posted by dgreenberger 7 months ago (Flag)
Thumb

Scott did mention that in the Google Document proposal, though it might be good to list it in the flowchart or otherwise provide a clear list of guidelines/caveats.

"One annoying caveat about POWO is that they don’t always seem to include the nominate infraspecies (e.g. Diplacus bigelovii var. bigelovii). Technically these should be created even though they are not in POWO if POWO does have non-nominate infraspecies in that species."

Posted by bouteloua 7 months ago (Flag)
Thumb

@bouteloua Thank you, I hadn't seen that doc!

Posted by dgreenberger 7 months ago (Flag)
Thumb

It would be nice to check the on-line Jepson for California specific plants otherwise we may end up without some subs and var.

Posted by efmer 7 months ago (Flag)
Thumb

As a way of thinking of 'disruptive changes' this implies, here's a quick analysis of the 'most observose' species in iNat that aren't accepted in POWO https://docs.google.com/spreadsheets/d/1JZ8HMRvRA6Wb7XPNgpLr-hGXn1SGNtX2UQlW5XxE3kU/edit?usp=sharing
I excluded names starting with P since those aren't very complete in POWO

So Anagallis arvensis has 5,473 obs and POWO suggests changing it to Lysimachia arvensis
Likewise, Heracleum maximum has 3,471 obs and and unclear why its not listed as accepted or a synonym in POWO, we should prob contact them about that one unless anyone sees an obvious explanation

Posted by loarie 7 months ago (Flag)
Thumb

@loarie What is going to be the protocol for contacting POWO about deviations etc.? I'm guessing a single point of contact at both ends, on a monthly (or other periodic) basis, would be preferable to constant random queries coming at them from multiple different directions (some possibly contradictory...).

Posted by jdmore 7 months ago (Flag)
Thumb

I've been seeing several fern taxon swaps based on POWO as the source...I'm not sure this workflow has gotten through to all. Should all ferns be flagged, or otherwise marked (to curators), or something else?

Posted by bouteloua 7 months ago (Flag)
Thumb

please tell whoever is swapping ferns not to use POWO as a source (along with P-genera) as described above. No one should be curating Ferns until there's clarity on what taxonomy we're using. See https://www.inaturalist.org/flags/271039

Posted by loarie 7 months ago (Flag)
Thumb

I was told this was the place to discuss issues with POWO. I'm trying to resolve https://www.inaturalist.org/flags/269917, which concerns what to do with forms of Potentilla tridentata, which is now known as Sibbaldiopsis tridentata on iNat. POWO does not include those forms (harking back to David's complaint above about spotty coverage of infraspecies), and to confuse matters, POWO seems to support both Sibbaldiopsis tridentata and Sibbaldia tridentata, which seems like an error. For my first issues of these forms, I walked through Scott's flow chart and added them to the "Vascular Plant POWO Deviations" doc. For the duplicate concepts in POWO, the chart suggests we should also have this duplication on iNat.

So @loarie, two questions I'd like you to answer (both of which seem like they've already been asked by David and Jim above but not answered):

1) How to we tell when POWO doesn't have a name because they reject it vs. when they have overlooked it and should be prompted to add it?
2) What is the process for getting POWO to update, e.g. resolving this issue of what seems like two active POWO names for the same concept? Should iNat curators email BI@kew.org directly, or are you perhaps monitoring the "Can't match" entries in the "Deviations" list and contacting Kew on a regular basis?

Posted by kueda 6 months ago (Flag)
Thumb

@kueda I just left a comment on the flag you referenced. Short version: support of both Sibbaldiopsis tridentata and Sibbaldia tridentata is definitely an error POWO needs to fix. And don't expect each and every name in IPNI to be accounted for in POWO. Many IPNI names are historical synonyms that are long out of use.

Posted by jdmore 6 months ago (Flag)
Thumb

1) How to we tell when POWO doesn't have a name because they reject it vs. when they have overlooked it and should be prompted to add it?

If POWO doesn't recognize the name as a synonym than its likely they've overlooked it. I'd recommend making and swapping into 'frankenstein' forms not in POWO (e.g. Sibbaldiopsis tridentata f. aurora) that match the POWO binomials

2) What is the process for getting POWO to update, e.g. resolving this issue of what seems like two active POWO names for the same concept? Should iNat curators email BI@kew.org directly, or are you perhaps monitoring the "Can't match" entries in the "Deviations" list and contacting Kew on a regular basis?

I just asked them which they'd prefer - standby. But in the meantime - the 'deviation' that we want to record is
that POWO has "Sibbaldiopsis tridentata" and "Sibbaldia tridentata" but iNat has just "Sibbaldiopsis tridentata"

Posted by loarie 6 months ago (Flag)
Thumb

Hi all,

This seems to be turning into us changing our 'no primary literature' policy, which is leading to some splits that have been really brutal in terms of time and communiy frustration. Is it time to step back and be a little less eager to rush into these things? I know in the past I've pushed for more scientific/'professional' functionality on inat and been told no, this isn't meant to be a hard science site, it's for connecting people with nature. Is thisstuff really helping people connect with nature? Sure seems to frustrate or confuse more people than it helps. In the least, if we are going this route can we be consistent in our approach so we can get more 'hard ecology' functionalty elsewhere on the site too? This really feels like the worst of both worlds.

Not trying to argue or be contrarian here... but I watched Mimulus aurantiacus get torn apart, first all accidentally put in the wrong species, then split but with most of them stuck in a section, causing tons of frustration and confusion. 'Amateur' people had no idea what was going on, IDers had to waste lots of time straightening it all out, and i had to deal witha bunch of changed IDs that blew up my notifications and took a lot of time i'd rather be spending on other things like helping people with IDs. Now we are talking about doing it with Spiranthes ceruna too, another 724 observations going to be chopped up. https://www.inaturalist.org/flags/269211 I am not questioning the research or hard work by taxonomists, but is this really how we want to do things? Can we take a step back here? please?

Posted by charlie 6 months ago (Flag)
Thumb

We are talking about a Spiranthes split, but it hasn't yet been committed and won't be committed until either a) the new taxa and ranges described by Matthew Pace are included in Plants of the World Online or b) there is rough community consensus* around the split.

People connect with nature in different ways. For some it's taking a walk in an area that has plants and animals and not just concrete, relieving stress, hearing birds, etc.. For others, it's observing, cataloging, and categorizing the components of those plants and animals and their communities. That includes taxonomy. Taxonomy includes splits.

*But yeah, what is rough consensus on iNaturalist and how is it achieved? (a question I raised 2 months ago on that long flag discussion).

Posted by bouteloua 6 months ago (Flag)
Thumb

I would at least agree that maybe there should be an enforced "gateway" consisting of a few of the most experienced taxonomic curators in each major group of organisms, through which changes affecting more than, say, a hundred existing observations would be vetted first before they can be committed. Vetting would consist of ensuring that (1) the proposed change or deviation has collected sufficient community consensus, and (2) it is being done in the least disruptive way possible (mechanically). Beyond that, their function would NOT be to agree or disagree with the scientific merit of the changes.

While I agree that taxonomic changes are inevitable and necessary, to the extent we can minimize their impact on connecting people with nature, it would be a good thing.

Posted by jdmore 6 months ago (Flag)
Thumb

I mean, i am not anti taxonomist or something. But this is something that is basically forced on everyone on the site not just those with interest in microsplitter taxonomy. Is the ecology behind it neat? sure. Do i want to spend hours of my time sorting through old Diplacus IDs i have made which are differentiated only really by range unless you have really specific pictures? Not at the top of my list but when we all adopt the taxonomy, we don't have much choice.

fwiw i don't see any reason why one couldn't continue collect ecological data without all the new splits. Things were working fine before. Not perfect, but fine. I think this sort of thing is, overall, way over done. Like, why aren't these all subspecies? Especially the ones that freely hybridize? But anyway. My point isn't to have the same rant again and again. I'm just throwing my hat int this 'rough consensus' and my vote is to tone the primary resource taxonomy down on iNat if not eliminate it completely (i did not realize POWO was stuffing in these primary literature changes so extensively or i might have voted no from the start). If I get outvoted, and i probably will given the rampant nature of splitters throughout taxonomy, so be it.

Posted by charlie 6 months ago (Flag)
Thumb

In place of my previous comment, I made this. All I can think of while reading this thread lol.

https://i.imgur.com/QgLMKTW.png

Posted by wdvanhem 6 months ago (Flag)
Thumb

hahahahahahaha

Posted by charlie 6 months ago (Flag)
Thumb

hahaha perfect

Posted by bouteloua 6 months ago (Flag)
Thumb

It's not that POWO entries are undisputed, some plants are moved so often, like Mimulus/Diplacus.
There are more recent papers suggesting that things are far from sorted out.
iNaturalist and many users disregard variations and subspecies, but the trend seems to be to remove variations and subspecies, invalidating hundreds of observations.
The automatic system is probably to blame, it's often amazingly good, but doesn't give the full name.
It's not always good to have the most recent names, sometimes older names work that much better in iNaturalist.

Posted by efmer 6 months ago (Flag)
Thumb

none of this is sorted out and extremely high numbers of name changes will continue for the next decade or two. Most people, including many if not most professional ecology types (in my experience at least) would rather just wait it out a while. There's no imperative to do so... if there's some special conservation issue we can address those directly.

Posted by charlie 5 months ago (Flag)
Thumb

The hardest push is going to come whenever new species are discovered and named. Sometimes they fit right into the existing taxonomic framework, but in other cases they force alteration of taxonomic concepts in related and/or parental taxa. If we wait, or shoe-horn those latter cases into an existing, now-inconsistent taxonomic framework, that only creates deferred maintenance that must eventually be addressed. And some iNat users are likely going to want to connect with those particular units of nature sooner than POWO's nominal 2-year incorporation process.

A case in point is https://scholarship.claremont.edu/aliso/vol30/iss1/6 (already noted in the Diplacus topic). Five obviously distinct species were named under Erythranthe. When those species were first added to iNat, Erythranthe had not yet been fully segregated from Mimulus in iNat. And the existing (still Mimulus) species under which they had formerly been lumped necessarily changed in scope too. How much other, analogous taxonomic inconsistency will we be willing to wait out? Will no one notice except taxonomists? (That's not a rhetorical question -- I'm open to being convinced!)

Bottom line, imperfect as it may be, I think the process that has been developed here is the best framework to address a perpetually thorny issue. (Crataegus taxonomy, anyone? ;-) Adoption of POWO necessarily forces us to address some of the deferred maintenance already accumulated. And doing that correctly necessarily means referring to the primary literature on which POWO is based.

(All of which is soooo much easier to do now than in the pre-online world of 30 years ago -- thus the accelerated rate of change.... sigh.)

Posted by jdmore 5 months ago (Flag)
Thumb

I mean, anything is based on primary literature if you go back far enough. Maybe i should just stick to the base issues.

-The process of splitting here doesn't work very well and cause mass scale problems and frustration especially if there isn't a geographic difference. If there is a geographic difference then we are mapping solely based on location which means we aren't actually learning anything about the species range or distribution because we aren't even looking.
-Taxonomy has become heavily dominated by splitters and any slight difference is described as a different species when it should be a subspecies (imho of course). Adopting this doesn't fit well with iNat's functionality.
-Drifting taxonomy has a very high cost in terms of database management, data collection, etc. both on here and elsewhere.

I understand that there are upsides to adopting the new names too, and i am not saying we shouldn't do it per se. I am saying i want the downsides to be recognized more fully, and for us to be open to different solutions other than 'everyone forced to accept the new splitting taxonomy with no effort to fix the issues' which is how this seems to be going now.

Instead of not adopting the changes we could also consider 'holding bin' taxa in the official taxonomy. In the past the iNat admins have been very resistant to that, but in this case it's a really simple (comparatively speaking) and effective solution that would stave off a ton of problems. Want to split Spranthes cernua into 12 species that all share the same range? Create a S. ceruna sensu latu designation and put it between genus and species. That way we don't have to adopt the new taxonomy or at least can do so gradually and have something to fall back on other than 'Spiranthes sp.' when we inevitably can't tell most of them apart.

And for goodness sake if we are going this route that involves catering to 'hard science' rather than the 'amateurs first' ideology that was pushed on us before, can we please consider adding some other ecology tools? Things like plot functionality, secret observations (even if you have to pay or something), negative data, improved mapping, improved geoprivacy, default fields, linking of observations, etc. I don't think it's a good approach to say 'iNaturalist is for amateurs and so doesn't support basic functionality for professional style monitoring, but we are forcing all our users to stay up to date with super obscure taxonomic changes'...

Posted by charlie 5 months ago (Flag)
Thumb

Where do I report things that are messed up.
Like Ambrosia salsola var. salsola.
There are 3 entries.
Ambrosia salsola salsola -> Ambrosia salsola NOT CORRECT
Ambrosia salsola salsola Correct
Ambrosia salsola salsola ??????

Posted by efmer 5 months ago (Flag)
Thumb

@jdmore Change isn't the problem, but loosing a lot of data and thus frustrating users is. They initially left out the variation not always a problem as in the same area there may be only one.
Next after the change the same plant is bumped back to the Genus level.
The conversion process is far from flawless, so we always check all our entries for the one or two the system somehow missed.

The biggest problem is a good on line reference that keys the new name.

Posted by efmer 5 months ago (Flag)
Thumb

Another one that drives you n....
Used to be Orobanche cooperi now is Aphyllon cooperi in the Jepson and Orobanche cooperi cooperi without a picture in Inaturalist.
So all valid Orobanche cooperi now seem to be invalid.
Why Orobanche cooperi cooperi ?????

Posted by efmer 5 months ago (Flag)
Thumb

@efmer Once upon a time O. cooperi consisted of ssp. cooperi and ssp. latiloba. The pair exist in iNat despite having been submerged into the parent in just about every other reference I can find. The Aphyllon switchover hopefully will be less awkward than what happened with Mimulus, although O. uniflora has been split (rightfully) into two occasionally sympatric species (A. uniflorum and A. purpureum) so that won't be fun

Posted by dgreenberger 5 months ago (Flag)
Thumb

@efmer if you find an issue with the taxonomy, like duplicate names, go to the taxon page for one of the species, or the genus, and
-click Curation
-Flag for curation
-leave a comment saying what the issue is

I fixed the duplicate Ambrosia salsola varieties.

Posted by bouteloua 5 months ago (Flag)
Thumb

@bouteloua Thanks, next time we will take that route, there are probably many more.

Posted by efmer 5 months ago (Flag)
Thumb

@dgreenbeerger Until we found this recent paper: http://www.phytoneuron.net/2015Phytoneuron/48PhytoN-Orobanchearizonica.pdf
So it it's wise to add a ruler to the flower shot.

Posted by efmer 5 months ago (Flag)
Thumb

These issues are complicated for sure, so I always think it's often prudent to use a stepwise approach if the taxonomy is less than "rough consensus"....just so you leave yourself with a chance to walk things back if you need to. Mike Hough and I have now looked at several 1000s of Spiranthes and it might be prudent to treat the Spiranthes cernua/incurva complex similarly to Nymphaea odorata complex. I get very queasy just using locations to assign species rank here, at least when your close to areas where overlap likely occurs.

Posted by mattyoung 5 months ago (Flag)
Thumb

@mattyoung That's why we need experts like you ready to zero in on the observations from the overlap areas and re-identify them if a split is implemented.

Posted by jdmore 5 months ago (Flag)
Thumb

Agreed -- some literally cannot be identified to species IMO though.....especially when looking at just a crappy photo or two. There are plants in the field that I've seen that look intermediate too, and so maybe as high as 20-25% can't be ID to species IMO.

Posted by mattyoung 5 months ago (Flag)
Thumb

And having them sit at Spiranthes (genus) or Diplacus sect. Diplacus, if necessary based on the evidence, is fine. There is an option in the Data Quality Assessment section where we can (carefully, measuredly) mark it as that the community cannot improve the identification.

Posted by bouteloua 5 months ago (Flag)
Thumb

Cool, good to know -- it's just we know the ID is basically between only "2 species" in the Spiranthes complex and it would seem you'd want some community agreement on how to ID them apart or at least be able to tag them as cernua/incurva.

Posted by mattyoung 5 months ago (Flag)
Thumb

@mattyoung if intermediate plants are frequent (more than just occasional spontaneous hybrids), then I would wonder what the justification is for treating them at species rank.

Posted by jdmore 5 months ago (Flag)
Thumb

exactly. i think the reason is that for some reason splitter papers are published more? Everyone wants to find a new species...

Posted by charlie 5 months ago (Flag)
Thumb

@charlie Not in my experience, at least. I've published new things at both variety and species rank, depending on what the patterns of variation (and other evidence) told me about gene flow (or lack thereof) between populations. But whatever the appropriate rank, there are definitely new things still out there to be discovered or recognized, and doing so doesn't automatically make someone a splitter. Otherwise Linnaeus would be the grandfather of all splitters ;-). And whether they get published or not is still up to the good ole peer review process.

Posted by jdmore 5 months ago (Flag)
Thumb

Curse Linnaeus the splitter.

I guess my main gripe is just with usability. If these super similar, intergrading things are to be considered species we need some formal subgenus type category to keep track of them. That somehow seems to get lost in the rush to classify each cryptic species but is crucial for broader ecology work as i'm sure you know. At least until we get handheld sequencers :)

Posted by charlie 5 months ago (Flag)
Thumb

We've done a first pass through all the "cernua" and "Spiranthes" records and ~20% can be assigned to a species with a high level of confidence. We came up with 45 cernua s.s. and 111 incurva out of 750-800 records. S. arcisepala is much more straight-forward though. And there's overlap, at least when phenotypically identified, in ranges of the species.

Posted by mattyoung 5 months ago (Flag)
Thumb

and wait til you get to the cacti..........

Posted by ellen5 3 months ago (Flag)
Thumb

Was recently going through observations of Veronica hederifolia/sublobata aka. V. hederifolia ssp. hederifolia/lucorum.

The inaturalist database currently has V. hederifolia and V. sublobata, as per POWO. But it then also has V. hederifolia ssp. hederifolia.

Distinguishing between the two taxa is tricky and unreliable, even with the specimen in hand, so in most cases, the suitable ID would usually be Veronica hederifolia sensu lato. It's not currently clear whether that's what V. hederifolia represents.

I've opened a flag: https://www.inaturalist.org/flags/314837

Posted by jobangles 2 months ago (Flag)
Thumb

@loarie How should I handle Malacothamnus to prevent a POWO lumping mess? As many of the Malacothamnus lumped by recent authors are considered rare taxa by CNPS, I'm considering them all valid until proven otherwise. My initial inclination is that all or most fit the taxonomy of Kearney which is what CNPS mostly follows and what was used until the 1993 Jepson. I'm just afraid someone will try to make the taxonomy fit POWO which would possibly require a lot of cleanup when things are probably mostly put back to the way they are currently on iNat.

Posted by keirmorse 2 months ago (Flag)
Thumb

You may want to flag the species/other taxa for curation and use the flags to discuss the taxonomy/propose deviations from POWO. For those trying to make the taxonomy fit POWO, there will then be a little text that says "Flagged for curation". This should indicate to those users that they should check the flags before committing anything. Here's an example with a summary of taxon changes (or not) in the first comment https://www.inaturalist.org/flags/276284

I started one on the genus here: https://www.inaturalist.org/flags/324799

Posted by bouteloua 2 months ago (Flag)
Thumb

Okay. Thanks. I'll see what I can do.

Posted by keirmorse 2 months ago (Flag)

Add a Comment

Sign In or Sign Up to add comments

Is this inappropriate, spam, or offensive? Add a Flag